Conventional network adaptation generally occurs either at the entire network level based on measurements sent to a network-level controller or at a local sub-network level without consideration of collateral effects on the overall network. As such, some implementations of conventional network adaptation technologies can result in complex analytical systems and much data transport to enable global network analysis, e.g., for centralized network adaptation schema. In other implementations of conventional network adaptation technologies, the local sub-network adaptation can impact the overall performance of the global network in undesirable ways. Alternative approaches to global network adaptation can become increasingly important as deployed network elements become increasingly heterogeneous and of finer grain in modern network topologies.